ID příspěvku ATH-3S7FHK
· Tier 1 papers
· Mathematical statistics
Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion) CC-BY
Archivovaná práce s otevřeným přístupem, importovaná pro dlouhodobé zachování (původ: Bayesian Analysis). Nejde o příspěvek na AthenXiv, proto nemá historii verzí ani historii časových razítek.
- Původně publikováno Jul 3, 2020 UTC
- Práva: CC-BY (datum vypršení není zaznamenáno)
- Přidáno na AthenXiv Oct 3, 2026 UTC
Nahrál jménem autora AthenXiv (UW2JTS24Y)
Abstrakt
Markov chain Monte Carlo is a key computational tool in Bayesian statistics, but it can be challenging to monitor the convergence of an iterative stochastic algorithm. In this paper we show that the convergence diagnostic Rˆ of Gelman and Rubin (1992) has serious flaws. Traditional Rˆ will fail to correctly diagnose convergence failures when the chain has a heavy tail or when the variance varies across the chains. In this paper we propose an alternative rank-based diagnostic that fixes these problems. We also introduce a collection of quantile-based local efficiency measures, along with a practical approach for computing Monte Carlo error estimates for quantiles. We suggest that common trace plots should be replaced with rank plots from multiple chains. Finally, we give recommendations for how these methods should be used in practice.
Klíčová slova: Markov Chains and Monte Carlo Methods, Statistics and Probability, Mathematics, Physical Sciences
DOI: 10.1214/20-ba1221
Licence: CC-BY
Číst online
Externí odkazy
- DOI Bayesian Analysis https://doi.org/10.1214/20-ba1221
Citovat tento příspěvek
Aki Vehtari, Andrew Gelman, Daniel Simpson, Bob Carpenter, Paul‐Christian Bürkner (2026).
Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion). AthenXiv.
https://athenxiv.com/paper/ATH-3S7FHK